Determination of Point-to-Point 3D Routing Algorithm Using LiDAR Data for Noise Prediction

Round 1

Reviewer 1 Report

Unfortunately, after carefully reading this article, nothing of interest was found.

Some observations about this work:

1. The objective is unclear since there are a lot the algorithms that compute the trajectory between two points. Thus, the main contribution is unclear, or a better explanation is missing.
2. The authors talk about noise sources, but it is confusing because practically all the elements in a scenario are considered noise. Then what is the real advantage? 
3. It would be a good idea to show the data from the LIDAR sensor and how the algorithm proposed is using it.
4. Please explain why you are comparing lidar against the algorithm proposed. Isn't your implementation supposed to use the data from that sensor?
5. It is imperative to compare the proposed algorithm facing others of the state-of-art.
6. In general, the whole paper needs a restructure to facilitate its compression and highlight the proposed methodology's advantages or inconveniences.

Author Response

Reviewer 1

Q1. The objective is unclear since there are a lot the algorithms that compute the trajectory between two points. Thus, the main contribution is unclear, or a better explanation is missing.

There are different algorithms for determination of shortest route between two points.  These routing algorithms primarily work in 2D environments. Largely the routing algorithms are used to determine the shortest route between two points in cities.  Typically, in a city environment all routes are required to be laid down between any pair of points before routing algorithm can determine the shortest route. Every route is defined by all the points or vertices (in X and Y) and the vectors joining the points making the individual routes. Routing algorithms attempt to determine the shortest of all established routes in 2D.   But the issue is that the existing algorithms primarily work when routes are delineated in 2D. Establishment of routes require digitization of terrain and route data. Initial abstraction of route is the prerequisite for any routing algorithm. Inaccuracy in delineation of routes causes ineffective determination of shortest route. Further, in case of sound propagation modelling, sounds propagate through principal routes following the physics of propagation. These shortest routes for sound propagation work in 3D similar to 2D shortest route discussed above.  For sound propagation modeling, actual propagation routes are not limited in 2D plane. In real scenario, From the source point there are many routes which are possible. But out of them there are certain shortest routes in different direction, which forms the set of principal routes. Existing Algorithms are not customized to the need of these computations. The computation requires the determination of all principal propagation routes from the source point to any specific destination point. Unlike the existing routing algorithm sound waves do not follow the network of previously laid down streets.    Route determination requires extraction of all routes in 3D between the source and destination points. Once the 3D routes are determined, the algorithm is required to find the shortest route for propagation of sound waves. In the proposed research a customized solution is evolved for extraction of routes between any pair of points (source and destination) and accurate shortest routes are determined in 3D using highly dense LiDAR terrain point cloud. The novel solution is applicable for sound propagation modelling in 3D and for any route determination applications in cities when previously labelled routes are not available. The paper not only developed a novel algorithm of 3D routing for sound propagation modelling. It further demonstrates the 3D noise mapping with the above algorithm. Thus, in summary the contributions of the paper are

1. Development of a novel algorithm to extract 3D shortest routes between a pair of points (source point and receiver point)
2. Determination of 3D routes using unlabeled raw 3D LiDAR terrain points existing between source and receiver points
3. Determination detailed routes with highest accuracy
4. Accurate sound propagation modelling integrating noise data, and LiDAR terrain data with noise model
5. As the technique extracts the routes in 3D accurately, it can very well improve the 2D routing algorithm used in different route determination algorithms

Q2. The authors talk about noise sources, but it is confusing because practically all the elements in a scenario are considered noise. Then what is the real advantage? 

 

Sound propagates from a point of high sound pressure level to point(s) of low sound pressure level. All points having higher sound pressure level work as the source points to their neighboring points having lower sound pressure levels. All points having certain sound pressure level can be a potential source for sound propagation. But out of these, those which are having significantly higher sound pressure levels compared to other are essentially playing the role of sources. From these source, noise/sound keep on propagating to all possible receiver locations where noise level is less. In real world outdoor environment sounds are received at a receiving location from several source points. Sound propagation modelling technique need to compute all such sound pressure levels receiving to a point to predict the sound pressure level accurately at a point for detailed noise mapping operation.

 

Q3. It would be a good idea to show the data from the LIDAR sensor and how the algorithm proposed is using it.

 

The Routing algorithm or the Noise propagation modeling essentially requires the terrain information as well as the noise information. So, LiDAR data is used to manage the terrain information. LiDAR points are unlike the other source for terrain information; i.e.,  it is more comprehensive, accurate, and 3D  . LiDAR comes in the form of point cloud (X, Y, Z). So, the challenge was? Can it be possible to develop an algorithm which can take LiDAR data directly and offer the principal routes for propagation of sounds and derive the accurate 3D noise map. Adapting LiDAR data directly would ensure high prediction accuracy compared with other approaches which take indirect terrain data such as Digital Elevation Model (DEM), Contour, Triangular Irregular Network (TIN), or rough height points etc., for route determination. The paper described in detail how LiDAR terrain data points can be used to extract building, ground, and other terrain points. It further determines all possible routes between any pair of points, and finally the shortest routes (principal routes) for propagation of sound from a source to  a receiver.

 

Q4. Please explain why you are comparing lidar against the algorithm proposed. Isn't your implementation supposed to use the data from that sensor?

 

The paper explains how LiDAR terrain data can be used for noise propagation modelling. The propagation modelling requires estimation of all the possible principal routes (shortest routes) between source and receiver points. For various noise source points every other surrounding points in the area work as receiving points (depending on relative difference of sound pressure levels). A point-to-point routing algorithm is developed which takes locations of sources and receivers and LiDAR terrain points in between. It determines the obstructions between source and receiving points. It ascertains the diffracting points, reflecting points, absorbing points and other transmission points to extract the shortest transmission routes in 3D between source and receiver pair.  Adapting LiDAR data directly had ensured high prediction accuracy compared with other approaches which take indirect terrain data such as Digital Elevation Model (DEM), Contour, Triangular Irregular Network (TIN), or rough height points etc., for route determination.

 

Q5. It is imperative to compare the proposed algorithm facing others of the state-of-art.

 

Various approaches used in route determination are explained in detail in the introduction part. Weaknesses in existing route determination algorithm are compared with the proposed route determination algorithm which can work in 3D and directly adapt to highly accurate LiDAR terrain point cloud data to determine the shortest routes for accurate sound/noise prediction. The advantage of proposed algorithm is further elaborated with the answer of Question no 1. Furthermore, the efficacy of new algorithm is established with detailed accuracy estimates for the novel algorithm, given under result and discussion section of the paper.

 

Q6. In general, the whole paper needs a restructure to facilitate its compression and highlight the proposed methodology's advantages or inconveniences.

 

The authors are thankful for the comments of the reviewer. Based on the suggestion of the reviewer the whole paper has been modified and restructured to become more concise. Further the proposed methodology’s advantages are highlighted in Section 2.

Reviewer 2 Report

In this text, a remarkable study on noise detection and 3D route estimation via LiDAR is presented. Although a comprehensive study has been carried out, some points need improvement. The language should be reviewed and improved. There is also a problem with article layout in general.  This situation reduces the intelligibility of the manuscript. My suggestions for improving the manuscript is following:
1) In the introduction, the relationship between noise and optimal route estimation should be explained in more detail with reference to the literature. The literature on the subject has not been adequately studied.
2) References to methods should be given in the appropriate place. for example; The Greedy algorithm… (Line 65), The Dijkstra algorithm (Line 69)
3) Move the first paragraph of Methodology to Introduction or Section 2. Or delete it completely.
4) Subdivide the “Methodology” into subheadings. Explain your data under the subtitle “LiDAR Data Used”. It makes the text easier to read.
5) Result and Discussion should be rearranged. Most of the formula information here should be found under the Methodology section. In the Result section, you should only present your results and make the necessary discussions. Method descriptions should not be in this section.
6) “Efficiency of algorithm” header should be removed. This section should be included in Result and Discussion.
7) In the Conclusion section, Line 714-721 should be in Result and Discussion. In addition, in the conclusion part, suggestions for the development of the study and future studies should be added.

Author Response

Reviewer 2

Q1. In the introduction, the relationship between noise and optimal route estimation should be explained in more detail with reference to the literature. The literature on the subject has not been adequately studied.

The authors are grateful for the comments of the reviewer and have acted upon them to improve the quality of the manuscript. The Literature part has been modified according to the comments of the reviewer and new literature have been introduced in the supplementary file of manuscript along with the introduction.

 

Q2. References to methods should be given in the appropriate place. for example; The Greedy algorithm… (Line 65), The Dijkstra algorithm (Line 69)

 

The authors thank the reviewer for their observation. The references have been cross-checked and all the errors have been rectified in the revised version.

 

Q3.  Move the first paragraph of Methodology to Introduction or Section 2. Or delete it completely.

 

The authors have read the manuscript again and updated the same accordingly.

 

 

Q4. Subdivide the “Methodology” into subheadings. Explain your data under the subtitle “LiDAR Data Used”. It makes the text easier to read.

 

The authors have updated the Methodology part and has also introduced the subtitles.

 

 

Q5. Result and Discussion should be rearranged. Most of the formula information here should be found under the Methodology section. In the Result section, you should only present your results and make the necessary discussions. Method descriptions should not be in this section.

 

The result and discussion section have been rearranged as per the recommendations. Method section has been modified accordingly.

 

Q6. “Efficiency of algorithm” header should be removed. This section should be included in Result and Discussion.

 

The authors have updated the manuscript according to the Reviewer’s comment.

 

Q7.  In the Conclusion section, Line 714-721 should be in Result and Discussion. In addition, in the conclusion part, suggestions for the development of the study and future studies should be added.

 

The authors have updated the manuscript according to the Reviewer’s comment. In conclusion part future scope is also introduced.

 

Scope for future study: The current study was focused on determination of point to point routing algorithm taking LiDAR data as terrain input.  The routing algorithm work in two stages, i.e., extraction of building, ground and other terrain points from LiDAR data in the first stage and then extraction of detailed routes between a source point and a receiving point in the second stage. The approach determined all the possible routes for propagation in 3D for every pair of source and receiver points and utilize it for sound propagation modelling integrating terrain and noise data with noise propagation model.  At present the extraction of building, ground and other terrain features between source and receiver points are accurate but semi-automatic in nature. In future the entire process is planned to be made automatic, thus algorithm will determine 3D shortest routes directly after receiving the LiDAR data, and noise data of noise sources for all the noise receiving locations. It would automatically, then, determine the 3D noise map for an area accurately.

Routing algorithm has important role to play in sound propagation modelling. Routing algorithms are important for various applications i.e., route planning, travelling salesman problem, laying out of pipeline, determination of sun angle, viewshed analysis etc. Automatic and efficient 3D routing algorithm working over detailed terrain data can improve the performance of all the above applications.   

 

Round 2

Reviewer 1 Report

After reading this new version, I realize that the authors have improved their work. Therefore, I think the paper is ready to be considered for publication.

Reviewer 2 Report

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

This paper employs LIDAR data to extract building, ground, and other major terrain parameters. The methodology is clear and specific, and the experimental setup is scientific and reasonable. Some minor issues are listed below:

Q1. There is an error in the format of equation 12, please be consistent with the format of equation 15.

Q2. Most references are not from the past 3 years, please follow up with the latest research results. References are not formatted uniformly.

Author Response

Q1. There is an error in the format of equation 12, please be consistent with the format of equation 15.

The authors are thankful for the comments of the reviewer. The formatting of the equations has been changed to keep a uniform format.

Q2. Most references are not from the past 3 years, please follow up with the latest research results. References are not formatted uniformly.

The formatting of the references have been updated and the older references have been attempted to be replaced with newer ones wherever possible.

Reviewer 2 Report

I have read the manuscript asi-1546184, entitled "Determination of Point-to-Point 3D Routing Algorithm Using LIDAR Data For Noise Prediction" written by Shruti Bharadwaj et al., submitted for publication and assessment in the journal ASI (ISSN 2571-5577).

Research presented in this manuscript is based on the use of TLS data as an environment for research of modelling and prediction of propagation noise. Derived results were compared with direct field measurements. The results are evaluated numerically and graphically.

A few remarks:

Manuscript has acceptable structure.

Abstract is concise and appropriate in scope.

The introduction section must provide a sufficient review of the literature as a basement for further work. The literature review is insufficient. Therefore, please expand the text of the section with as many relevant sources of cited literature as possible.  Sources should be cited accurately and directly.

There is a need to improve the quality and readability of images throughout the manuscript.

Add the localization image in section 3.

Section 4.1. must be significantly expanded. Complete the field measurement methodology, TLS data processing method, georeferencing, quality and accuracy of point clouds, etc.

The results of the research are presented numerically and graphically in Chapter 4. However, authors present very many details of processing that in my opinion do not belong to the text of the manuscript. E.g. line 310, 317, 327, 329, 339, 351, 379, 386, and so on, etc. If authors need to provide this and similar information, I recommend that they be included in a separate Appendix section.

Add units on axes in the figures and graphs.

In the Discussion section, there are no citations. Authors should discuss their results with the results of other authors and researches on similar types of objects. A lot of published papers correspond to the purpose of this research.

Results must be reported more clearly. They must be sought in this version of the manuscript.

Some sources in the reference list do not conform to the MDPI format - please edit. References including conference outputs - I recommend authors to use more references from journals and expand the list of references (maybe double them).

Some minor comments:

Check the text and correct small editing errors: e.g. Line 47, 56, 67, 73, etc. Correct in the text LIDAR, LiDAR, Lidar.

Overall:
The manuscript after these corrections can be accepted for the next publishing process. 

Author Response

 

The introduction section must provide a sufficient review of the literature as a basement for further work. The literature review is insufficient. Therefore, please expand the text of the section with as many relevant sources of cited literature as possible.  Sources should be cited accurately and directly.

The authors are grateful for the comments of the reviewer and have acted upon them in attempt to improve the quality of the manuscript. The introduction has been modified according to the comments of the reviewer and new references have been introduced in the manuscript.

There is a need to improve the quality and readability of images throughout the manuscript.

The quality of the figures have been attempted to improve wherever possible

Add the localization image in section 3.

The localization image has been added as Figure 5 in the section 3 of the manuscript.

Section 4.1. must be significantly expanded. Complete the field measurement methodology, TLS data processing method, georeferencing, quality and accuracy of point clouds, etc.

The section 4.1 has been significantly modified and appropriate content is added in the manuscript

The results of the research are presented numerically and graphically in Chapter 4. However, authors present very many details of processing that in my opinion do not belong to the text of the manuscript. E.g. line 310, 317, 327, 329, 339, 351, 379, 386, and so on, etc. If authors need to provide this and similar information, I recommend that they be included in a separate Appendix section.

The section 4 in the manuscript has been modified to include only the significant data

Add units on axes in the figures and graphs.

Units and axes have been added on the graph wherever required

In the Discussion section, there are no citations. Authors should discuss their results with the results of other authors and researches on similar types of objects. A lot of published papers correspond to the purpose of this research.

Citations have been added in the discussion section based on the similar types of research that has already been publoshed

Results must be reported more clearly. They must be sought in this version of the manuscript.

The results section has been modifed

Some sources in the reference list do not conform to the MDPI format - please edit. References including conference outputs - I recommend authors to use more references from journals and expand the list of references (maybe double them).

The formatting of the bibliography has been modified to be more uniform than earlier

Some minor comments:

Check the text and correct small editing errors: e.g. Line 47, 56, 67, 73, etc. Correct in the text LIDAR, LiDAR, Lidar.

The occurrences of the word has been checked and made uniform throughout the manuscript.

Overall:
The manuscript after these corrections can be accepted for the next publishing process. 

The authors would once again wish to thank the reviewer for their constructive comments

Reviewer 3 Report

In this paper, the authors develop algorithm extracts building, ground, and other terrain parameters by employing a 3D point cloud. These terrain parameters are utilized further for the point-to-point detail routes determination for the noise propagation by employing the novel cutting plane method because it helps in deriving the different routes quite accurately. Moreover, this algorithm integrates the terrain data and the noise data for the mapping of noise levels in 3D using a sophisticated noise model. This paper has some innovation and abundant references, but there are still some problems to be noted.

 

(1) Abstract aims to describe the purpose, method and final conclusion of the study briefly. I think the content of the abstract is too much. Do you consider simplifying and deleting some redundant content?

 

(2) There are two parts missing at the end of your introduction. One is the list of contributions to the article. The other is the introduction of the following parts of the article. I hope you can add relevant content for readers to read and understand.

 

(3) I noticed that the text in your pictures seems to have obvious distortion, and we can't see it clearly even after magnifying. Could you please adjust the size and clarity of the picture?

 

(4) It seems that you have not explained the formula in detail. For example, I do not know the respective meanings of x, x1, x2 and y, y1 and y2 in equation 1,. Would you consider adding a relevant description?

 

(5) There seem to not exist the compared algorithms. If the authors do not compare with some state-of-the-art methods, how to completely validate the performance of the proposed method?

Author Response

(1) Abstract aims to describe the purpose, method and final conclusion of the study briefly. I think the content of the abstract is too much. Do you consider simplifying and deleting some redundant content?

 The authors are thankful for the comments of the reviewer. Based on the suggestion of the reviewer the abstract has been modified to reduce the redundancies and become more concise

(2) There are two parts missing at the end of your introduction. One is the list of contributions to the article. The other is the introduction of the following parts of the article. I hope you can add relevant content for readers to read and understand.

The authors have modified the introduction of the article and have also added references that are clearer and provide more insight to the readers. The introduction in the following parts of the articles has also been added. Since it is also a requirement to keep the article concise and short, the additions have been subsequently managed to follow that requirement as well.

(3) I noticed that the text in your pictures seems to have obvious distortion, and we can't see it clearly even after magnifying. Could you please adjust the size and clarity of the picture?

The quality of the images have been modified wherever possible.

(4) It seems that you have not explained the formula in detail. For example, I do not know the respective meanings of x, x1, x2 and y, y1 and y2 in equation 1,. Would you consider adding a relevant description?

The relevant descriptions have been added in the manuscript wherever required

(5) There seem to not exist the compared algorithms. If the authors do not compare with some state-of-the-art methods, how to completely validate the performance of the proposed method?

The authors are thankful for the comments of the reviewer but with all due respect, the authors want to say that the comparison part is mentioned in the introduction section as well as in the section where the need to generate an algorithm is written. Further, the algorithm is validated by a two-way accuracy check. Firstly, the accuracy of routes (route over the top, route around the sides, and reflected route) is validated mentioned in Figure 23 and Figure 24. This route accuracy measurement generates a point-to-point accurate 3D route. Secondly, the validation is done for the noise prediction measured by 3D routing. For the validation, short-term instantaneous noise prediction is performed which gives a max error in between 6-6.5 dB. While the traditional algorithm shows the error in between 10-12 dB (referenced in the conclusion text).

 

Round 2

Reviewer 2 Report

Thanks to the authors for their responses to the comments.

Authors made significant improvements with content and formal changes in the manuscript. 
Recommendations resulting from the first round of review were responded and/or corrected in the manuscript.

My previous recommendation that there is a need to improve the quality and readability of figures throughout the manuscript still persists.

After adjustment, I recommend this version of the manuscript for further publication process.

Author Response

• Authors made significant improvements with content and formal changes in the manuscript. 
  Recommendations resulting from the first round of review were responded and/or corrected in the manuscript.

The authors are glad that the reviewer found the revisions to be satisfying and would like to thank the reviewer for all the comments that enriched the quality of the manuscript.

• My previous recommendation that there is a need to improve the quality and readability of figures throughout the manuscript still persists.

The authors have made sure to enhance the quality of all the figures and make them more readable. The authors hope that the reviewer will find the modified figures to be up to the mark.

• After adjustment, I recommend this version of the manuscript for further publication process.

The authors once again thank the reviewer for all their past comments and are glad that the revised version has been recommender for publication.

Reviewer 3 Report

In this paper, the authors develop an algorithm to determine all the possible routes for propagation, beginning with 3D LIDAR point cloud data. The developed algorithm extracts building, ground, and other terrain parameters by employing a 3D point cloud. These terrain parameters are utilized further for the point-to-point detail route determination for the noise propagation by employing the novel cutting plane method. It helps in deriving the different routes quite accurately. The content of this article is innovative to some extent, but there are some problems, hope to attract the authors’ attention.

(1) From the reviewer's point of view, reading the revision mode you uploaded is very laborious and inconvenient. Can you re-upload the manuscript in the normal format next time you submit it?

(2) Does the deleted in schema mean a complete deletion or a modification? I would like you to re-condense the abstract and conclusion, focusing on the work of this paper and refining the background description.

(3) In the introduction, it is suggested that the authors clearly list the contribution points of this article according to the serial number, so as to improve the readers' interest in reading and make readers better understand the article.

(4) Figure 6 and Figure 7 are not exquisite, and the square frames at the same level are not aligned. I think you have the ability to draw a better flow chart, and I hope you can revise and improve them.

(5) This article puts forward a good route determination method. I'm curious about your next work plan, how to improve this algorithm and its application scenario expansion. Can you add relevant content?

Author Response

• From the reviewer's point of view, reading the revision mode you uploaded is very laborious and inconvenient. Can you re-upload the manuscript in the normal format next time you submit it?

The authors are apologetic for the inconvenience caused while reading. The revised manuscript has been uploaded by adhering to the normal format and the authors assure that reading the revised manuscript will be convenient.

• Does the deleted in schema mean a complete deletion or a modification? I would like you to re-condense the abstract and conclusion, focusing on the work of this paper and refining the background description.

In the revised version, there is no complete deletion. The authors have instead made the necessary modifications. Adhering to the reviewer’s comment, the authors have made the necessary modifications in the abstract and conclusion as well. In this, the focus of the paper and a description of the background has also been mentioned.

 

• In the introduction, it is suggested that the authors clearly list the contribution points of this article according to the serial number, so as to improve the readers' interest in reading and make readers better understand the article.

In the revised article, the authors have followed the reviewer’s recommendation and have listed the contribution points of the article based on the serial number.

(4) Figure 6 and Figure 7 are not exquisite, and the square frames at the same level are not aligned. I think you have the ability to draw a better flow chart, and I hope you can revise and improve them.

The authors understand the reviewer’s concerns and have made significant modifications in the flow chart. The authors hope that the reviewer will like the new flow chart.

 

(5) This article puts forward a good route determination method. I'm curious about your next work plan, how to improve this algorithm and its application scenario expansion. Can you add relevant content?

The authors are thankful to the reviewer for the appreciation. In the conclusion section, the authors have mentioned about our next plan of action, ways to improve the algorithm and also the fields in which this work can be applied. The authors hope that the reviewer will find the revised version of the manuscript to be good enough.

Round 3

Reviewer 3 Report

In this paper, authors propose an algorithm, which uses the novel cutting plane technique customized to work with LiDAR data to extract all the principal routes between every pair of noise source and destination. Once the routes are determined these are used to determine important terrain parameters for acoustic modeling The terrain parameters, and noise data when integrated with sophisticated noise model gives accurate prediction of noise for a place. The point-to-point route determination algorithm is developed using LiDAR data of RGIPT campus. This article is innovative to some extent, but the details of the article and the opinions put forward last time are still not perfect, so it is not recommended to publish.

(1) You have revised the abstract, but it is not concise enough. I hope you can focus on the contribution of this paper and reduce the introduction of background and research status.

(2) In order to make it easier for readers to understand, I suggest you list the main contribution of the paper according to the serial number in the introduction, but I didn't find that.

(3) The distortion of the figure text is not only caused by the font, the proportion of the figure is the main reason. I hope you can pay attention to the details, improve the aesthetic degree.

(4) The beauty of your modified picture has been greatly improved, but why do you not pay attention to the details? For example, the lines in the flowchart in the second box are not on the same line.

(5) I noticed that the figures you use have similar formats and contain text. Since you have changed the text font in Figure 2, why not modify the text format in other figures at the same time?